The recent explosion in the number of novel macromolecules, many identified by the genome sequencing efforts, has intensified the need for improved compositions and methods for separating macromolecules. Thousands of recently identified macromolecules have yet to be purified and characterized functionally. Techniques for the rapid capture, isolation, detection, analysis, and quantification of macromolecules would accelerate the functional characterization of novel macromolecules.
In particular, when the macromolecule is a protein, current methods of capture and separation are cumbersome and expensive. In one current technique, affinity matrices are used to capture and/or separate a protein of interest from a mixture of proteins and other molecules. Affinity matrices might be prepared using a purified sample of the protein to create antibodies. However, the preparation of antibodies that specifically bind a protein can take several months and might even require a purified sample of the protein. Alternatively, an affinity matrix might be prepared using knowledge of the function of the macromolecule. For instance, an affinity matrix based on a binding partner of the protein might be used for capture and separation of the protein. Unfortunately, methods of separating macromolecules that require extensive knowledge about the macromolecule, or even a purified sample of the macromolecule, are ineffective with macromolecules that have yet to be characterized. The requirement of a purified sample of the macromolecule for the preparation or selection of an affinity matrix often presents researchers with nothing more than frustrating circularity.
In addition to affinity matrices, other techniques are also used to separate macromolecules. For example, the current state of the art technique for separating large numbers of proteins is two-dimensional gel electrophoresis. The technique typically resolves about 2,000 proteins, and the best gels can resolve up to 11,000 proteins (Abbot, 1999, Nature 402:715-720). Unfortunately, many researchers require separation techniques that can resolve proteins from samples as diverse as the entire protein fraction of a mammalian cell. The protein fraction of a cell can contain tens of thousands of proteins, overwhelming the resolving power of 2D electrophoresis (Abbot, 1999, supra). Since the full sequences of the genomes of many species, including humans, are nearing completion, researchers must now grapple with the functions of hundreds of thousands of novel macromolecules (Abbot, 1999, supra). New techniques of macromolecular separation that require limited information or even no information about the target macromolecules are needed.
Currently, researchers are seeking improvements in protein separation. For instance, some are attempting to create chips which specifically bind proteins. In one typical chip, antibodies specific for known proteins are attached to a substrate to form a microarray. These chips can then be used to bind and identify proteins from a complex solution (Abbot, 1999, supra). However, these chips suffer from the same limitations of antibody production that plague affinity matrices. For each protein to be bound by the chip, a unique antibody must be prepared by an expensive process that can take several months. In addition, many proteins are not sufficiently immunogenic to create antibodies for binding.
In the field of small molecules, the technique of molecular imprinting has provided an efficient method for the preparation of matrices that are capable of selectively binding a target molecule. To prepare a molecular imprint, a matrix is formed around a template molecule. After the matrix has formed and the template molecule is removed, the matrix can then be used to selectively bind the template molecule. As early as 1949, a silica gel was created that selectively bound a dye (Dickey, 1949, Proc. Natl. Acad. Sci. USA 35:227-229). Recently, an imprint prepared with phenyl-α-D-mannopyranoside was sufficiently selective to resolve a racemic mixture of the saccharide (Wulff, 1998, supra).
Current methods form imprints of molecules in organic polymers (Wulff, 1998, Chemtech 28:19-26). To create cavities of defined shape, polymerizable molecules are bound, covalently or noncovalently, to a template molecule (Wulff, 1998, supra). The resulting complex is copolymerized in the presence of a large amount of a cross-linking reagent (Wulff, 1998, supra). The templates are then removed leaving microcavities with defined shapes and arrangements of functional groups (Wulff, 1998, supra). Imprints made by such a technique display selective binding for the template molecule. Molecular imprints have been used for chromatographic separation, immunoassays, chemosensors, and even catalysis (Wulff, 1998, supra).
To date, molecular imprints have had limited application to the binding of larger molecules including macromolecules. In fact, one review states that only small molecules can be imprinted with any great confidence. Molecular imprints of larger molecules like nucleic acids, peptides, proteins and cells fail because larger molecules yield more heterogeneous binding sites and because larger molecules can be too fragile for conventional methods of molecular imprinting (Cormack and Mosbach, 1999, Reactive and Functional Polymers 41:115-124).
Nevertheless, a few successful imprints of larger molecules have been produced. Synthetic polymers which selectively bind amino acid derivatives and peptides were created using the target amino acid derivative or peptide as a template (Kemp, 1996, Anal. Chem. 68:1948-1953). Imprints have also been created which bind to nucleotide derivatives (Spivak and Shea, 1998, Macromolecules 31:2160-2165). Ionic molecular images of polypeptides have been created by mixing a matrix material with the intact polypeptide chain to be bound by the molecular image (U.S. Pat. No. 5,756,717). Molecular imprints of cytochrome c, hemoglobin and myoglobin, respectively, have been prepared by polymerizing acrylarnide in the presence of each intact protein (U.S. Pat. No. 5,814,223). An imprint of horse myoglobin selectively bound horse myoglobin from a mixture of proteins including whale myoglobin (U.S. Pat. No. 5,814,223).
Although current methods of molecular imprinting have shown limited initial success at selectively binding macromolecules, the current methods are not sufficient for the efficient capture of macromolecules. The current techniques for molecular imprinting require a purified sample of the macromolecule to be bound by the imprint. The inability to produce a specific imprint in the absence of a purified sample of the macromolecule is no different from one of the failings of conventional methods of protein separation. In addition, current methods of preparing molecular imprinting are not amenable to creating the thousands of imprints often required by current large-scale experiments. Purification of hundreds or thousands of proteins to create a matrix for separating the proteins of a cell extract is no more efficient that 2D electrophoresis. An efficient method for producing compositions that selectively bind macromolecules given limited information about the structure or function of those macromolecules is needed. An ideal method could produce a composition capable of binding a macromolecule given as little information as a partial primary structure of the macromolecule.
An improvement in molecular imprinting to enable the preparation of affinity matrices in the absence of a purified sample of the macromolecule would overcome many limitations of the art of molecular imprinting. Techniques are also needed to efficiently separate and identify thousands of proteins. Compositions with specificity for macromolecules that can be produced rapidly and at low cost will enable such techniques. Ideal compositions would be arrays of such binding compositions, each composition designed to bind a given macromolecule. Such an array could be used to rapidly screen a complex biological sample for a number of different macromolecules simultaneously.